Computational parameters and neural dynamics of state representation processes to parse pathophysiology of early psychosis

University Of Minnesota, Minneapolis MN

Investigators

Linked publications & trials

Abstract

PROJECT 4 - SUMMARY PROJECT 4 provides the translational link between our Center’s pre-clinical research on behavioral and neural manifestations of disrupted state representation processes and their clinical consequences in people with early psychosis. We will do this by examining EEG and MRI signals associated with state representation changes across a broad spectrum of early psychosis patients and community controls. All participants complete the cognitive control task (TOPX) and reward-based decision-making task (TBT) being used across our Center that allow modeling of state estimation, maintenance, and learning. We will use the parameters derived from the computational models to investigate MRI and EEG indices in the central executive network (CEN), the salience network (SN) and the default mode network (DMN), and to examine clinical symptoms and patient subprofiles that derive from different patterns of disruption. Our overarching goal is to test the utility of computational modeling of state representation disruptions in parsing the heterogeneity of early psychosis to align with pathophysiological mechanisms, with the aim of informing future precision psychiatry interventions. In Study 1, we continue to analyze longitudinal and simultaneously acquired fMRI/EEG in 100 early psychosis (EP) patients and 100 community controls. In Study 2, we add a new EP sample (n=150) with matched controls (n=150) using high-density EEG during extended versions of our two Center tasks to replicate and extend upon our recent findings of distinct computational subprofiles in EP individuals. In Study 3, we collect a large online sample (n=1,000) to optimize the efficiency of computational assessment across the two tasks. Aim 1: Characterize state representation disruptions in psychosis using computational modeling of a cognitive control task and a reward-based decision-making task. We will model trial-level behavioral data to test the reliability and generalizability of state representation processes assessed via TOPX and TBT. Aim 2: Determine neural functions associated with state representation disruptions in early psychosis. We will map state representation parameters onto neural functions, using EEG-fMRI data acquired during the current cycle, and in an independent sample of EP individuals via EEG metrics measured more precisely. Our focus will be on characterizing control dynamics of CEN, SN, and DMN interactions. Aim 3: Validate dimensions of computational variation in early psychosis and determine their clinical, cognitive, and neurophysiologic correlates. With data from all three studies, we will perform classification and information theoretic analyses of state representation parameters, in order to see if we replicate the computational subprofiles identified in the first funding period and in order to extend upon their relation with clinical and neurophysiological measures.